The relative contributions of impaired insulin secretion and of tissue insensitivity to insulin to the carbohydrate intolerance of uremia were investigated in 10 chronically uremic subjects. Two types of glucose-clamp experiments were performed in each patient before and after 10 wk of thrice weekly hemodialysis. In both types the blood glucose concentration was maintained at a constant level by the periodic adjustment of a variable glucose infusion with a negative feedback formula.Hyperglycemic clamp. The blood glucose concentration was acutely raised and maintained 125 mg/dl above basal levels for 2 h. Since the glucose concentration was held constant, the glucose infusion rate is an index of glucose metabolism (M). After dialysis M increased in all patients from an average of 4.23 to 6.30 mg/kg body wt per min (P < 0.001). The plasma insulin responses (I) both pre- and postdialysis were biphasic with an early burst within the first 2-5 min, followed by a phase of gradually increasing insulin concentration. After dialysis the plasma insulin response diminished slightly. Consequently, the M/I ratio, an index of tissue sensitivity to endogenous insulin, increased postdialysis in all subjects by an average of 92% (P < 0.01). Euglycemic clamp. The plasma insulin concentration was acutely raised and maintained by a primecontinuous insulin infusion. The blood glucose concentration was held constant at the basal level by a variable glucose infusion as above. M/I again is a measure of tissue sensitivity to insulin (exogenous) and increased in all patients postdialysis by an average of 57% (P < 0.01). In two patients hepatic glucose production was measured with tritiated glucose during the euglycemic clamp and declined by 84% predialysis. A similar decrease (82%) was observed postdialysis. Thus, both the hyperglycemic and euglycemic clamp techniques demonstrated tissue insensitivity to insulin to be the dominant carbohydrate defect in uremia. The surprising apparent lack of consistency in the change in beta cell response postdialysis is explained by the strong inverse correlation between beta cell sensitivity to glucose and tissue sensitivity to insulin (r = -0.920; P < 0.001). Those individuals who showed the most striking improvement in tissue sensitivity to insulin actually decreased their serum insulin response to hyperglycemia; those whose improvement in tissue sensitivity was more modest showed increases in beta cell responses.